2. Business IJBGM Exploring Board Size And Firm Performance Bongo Adi Nigeria

0.47 MB·English

International Journal of Business and
General Management (IJBGM)
ISSN(P): 2319-2267; ISSN(E): 2319-2275
Vol. 3, Issue 1, Jan 2014, 9-30

© IASET

EXPLORING BOARD SIZE AND FIRM PERFORMANCE: IS THERE A KUZNETS’
CURVE RELATIONSHIP? EVIDENCE FROM NIGERIA
CHRIS OGBECHIE & BONGO ADI
Lagos Business School, Pan-Atlantic University, Lagos, Nigeria

ABSTRACT
Board of directors is an essential component of corporate governance especially in the post-crisis global
organizational system. The Agency theory and the Resource Dependency theory emphasize the role of board size in
determining the effectiveness of boards in corporate governance. This study draws from both theories and considers the
issue of board size and organizational performance not from the perspective of how large or small boards should be, but
from the angle of optimality. In so doing, the study explores the existence of an optimal board size for firms in Nigeria, as
well as the impact of board size on organizational performance.
Using both mean reversion trend and random effect model the study finds the existence of a Kuznets’ curve in the
relationship between board size and firm performance and in line with previous studies but no optimal board size for the
sampled 100 firms in Nigeria. However, when this sample is divided into financial and non-financial firms, we find
evidence that the target board size is at least 12 directors for the financial firms. Estimates of the Arrelano-Bond dynamic
panel data estimation further show that board size does not significantly influence organizational performance.
KEYWORDS: Firm Performance, Resource, Kuznets, Relationship, Nigeria
1. INTRODUCTION
The idea of a dichotomy between share holders objective and that of firms' managers became predominant in the
1960s. The view then is that managers would seek to maximize their own utility first then consider the implications of this
for the firm (Baumol, 1959, 1962; Marris, 1964; Williamson, 1963). This position closely approximates the views of
Berle & Means (1932) who recognized that the separation of shareholders' ownership and management control in a
corporation results in agency problem The agency problem arises when the principal (shareholders) lacks the necessary
.
power or information to monitor and control the agent (managers). More so, the shareholders, in most cases, comprising
many individuals in diverse locations will always be at the mercy of managers when urgent decisions must be taken; they
will find it difficult reaching a consensus in such circumstances even if owners’ negotiation was initiated.
The above scenario made it critical to have some control and oversight over managers. Company boards,
therefore, became an essential component of corporate governance, with the role of monitoring management
(Shleifer & Vishny, 1997). Recent discourseon corporate governance has overtly emphasized the role of company boards
in organizations' performance. For instance, it is believed that a good board can provide a link between the firm and its
environment, secure critical resources (Hillman et al., 2000).
It has an important role to play in helping management make strategic decisions (Davies, 1999; Kemp, 2006).
Furthermore, an effective board is likely to help the firm achieve superior performance (Hawkings, 1997). And is a source
of strength in several ways such as attracting investment capital, improving valuation and share price performance and
providing better long-term shareholder returns (Lee, 2001; Carlsson, 2001).
10 Chris Ogbechie & Bongo Adi

Despite the above emphasis on boards’ positive contribution to firm performance, much of the recent corporate
failures have yet been linked to company boards. Ogbechie (2012) notes that the financial meltdown of 2008 that led to the
collapse of companies warranted more focus on board effectiveness, leading to emphasis on transparency, accountability,
regulatory oversight and efficiency of boards. Again, the near collapse of 10 banks in Nigeria in 2009 was also connected
to the activities of boards, leading to the dismissal and prosecution of boards of directors of 8 banks in the country.
Even recent corporate scandals in corporations such as Enron, WorldCom, Tyco international in the U.S., HIH insurance in
Australia, Parmalat in Italy, Saytem in India highlight the inadequate role played by boards, leading to patent company
failures (France & Carney, 2002; Weekend Herald, 2003; Economist, 2003; Lockhart, 2004; Ogbechie, 2012).
These re-occurring corporate failures has spurted interest on what makes for an effective board. Researchers and
Policy makers’ focus have been on specific board characteristics – outsiders' representation, board size, board leadership,
Board diversity1 and CEO duality2– in explaining boards’ effectiveness. In some of the studies, there has also been
emphasis on factors such as board process, board committees and director characteristics. But in all these, board size in
developing markets has not been adequately treated, despite evidence in the theory that board size is critical for board
effectiveness, nay organizations’ performance. The agency theory and resource dependency theory for instance support the
notion that board size has significant economic impact on firm value (Ning et al., 2010; Eyenubo, 2013). While the agency
theory argues that large boards can be costly, increasing operational complexity and the potential for dissention among
members and leading to lower organization value (Lipton & Lorsch, 1992; Jensen, 1993; Eisenberg et al., 1998), resource
dependency theory suggests that companies are better off with large boards – each new board member brings both
expertise and access to resources, and the diversity allows the board members to provide management with high quality
service (Zahra & Pearce, 1989). So, what actually offers the optimal board size, the agency theory or the resource
dependency theory?
Indeed, large boards are susceptible to social loafing. They also “tend to be contentious and fragmented, which
would reduce their ability to monitor and discipline senior management. In such cases, the senior managers can
affirmatively take advantage of the board through coalition building, selective channeling of information, and dividing and
conquering (Alexander et al., 1993).But that does not make smaller board size more appealing, for in this case there are
issues of group think and limited insight. So, it boils down to answering the question, “how large or small should
companies’ boards be?”
Various suggestions as well as institutional standards have been established to guide board size in various climes.
In Nigeria, the code of Best Practices for Public Companies recommends that the Board of Directors should be composed
of both executive and non-executive directors and should not exceed 15 persons or be less than 5 persons in total.
The Code of Corporate Governance for Banks in Nigeria on the other hand recommends a board size of between 5 and 20
directors. So, between these wide margins, what constitutes an optimal board size still remains uncertain.
Sanda et al (2010) using a sample of 93 firms for the period 1996 through 1999 in Nigeria proposed an optimal board size
of 10. Some other scholars have argued that board size should be no greater than 8 or 9 for all firms (Lipton &
Lorsch, 1992; Jensen, 1993). Hilb (2004) on the other hand recommends a small, legally accountable, and well-diversified
board comprising a maximum of 7 members (including an independent chairman, independent members and the CEO) as

1Board diversity refers to the make-up of a company’s board. It implies that the directors have different skills, knowledge
and experience should be from different age groups and social status.Diversity can either be observable, such as gender,
age, race and ethnicity, or non-observable such as knowledge, education, values, perception, affection and personality
(Ogbechie, 2012).
2CEO duality is a situation where a firm’s CEO also serves as the chairman of the board of directors.
Exploring Board Size and Firm Performance: Is there a Kuznets’ Curve Relationship? Evidence from Nigeria 11
the ideal board size for publicly quoted firms. And Ning et al. (2010) proposes an optimal board size of 8–11 directors, as
U.S. publicly traded corporations tend to target this number over time. Thus there is yet to be a consensus on the issue of
optimal board size.
The divergent views are believed to result from the difference in institutional, geographical and firm specific
factors which in fluence board size decision. Such differences, which are very obvious between the developed and the
developing markets, warrant a view of board size optimality from the lens of the developing markets. As such, in view of
the little attention given to board size optimality in Africa, and in Nigeria specifically, this study will make significant
contributions to the literature as well as corporate governance mechanisms with the following objectives: 1) to explore
what constitutes optimal Board Size for companies in Nigeria.2) To examine the significance of board size for companies’
performance, while controlling for endogeneity problems. The paper proceeds as follows: Section 2 provides a review of
the literature on board Size. Section 3 explains the methodology and discusses the dataset used. Section 4 presents the
empirical results, discusses them and relates them to the different strands of the literature. Finally, Section 5 concludes the
paper.
2. BOARD SIZE OPTIMALITY: LITERATURE REVIEW
The debate on how large or small a corporate board should be has dominated the interest of researchers. This is
not surprising, because of the substantial role boards have to play in corporate governance, and the understanding that size
matters for the boards’ efficiency. Two rival theories – Agency theory and Resource dependency theory – have emerged
over the years as explanations to the question, “how large should board size be?”
Agency Theory
The theory argues that as firms grow in size, the shareholders (principals) lose effective control, leaving
professional managers (agents) with specialized knowledge to manage the affairs of the business, and in so doing, these
agents over time gradually gain effective control over the business (Mizruchi, 1983). Often times, this transfer of firm’s
control from principals to agents over time leaves the agents free to pursue their own selfish aims to the detriment of the
principals. This situation provides potential for managerial mischief, which leads to the agency problem
(Dalton et al., 2007). As such, some internal and external control mechanisms that will help check the excesses of the
agents were required, and board of directors, became the suitable body to control the corporate agency problem, by
governing management’s decisions and assessing their impact on shareholders’ wealth (Alchian & Demsetz, 1972;
Fama & Jensen, 1983).
Agency theory defines the monitoring effectiveness of the board in terms of size and independence
(Ogbechie, 2012). The proponents, as well as various empirical studies, argue that a substantial increase of the board size
could result in a slowdown in decision making and an increase in costs, leading to lower firm/organizational performance
(Alexander et al., 1993; Yermack, 1996; Eisenberg et al., 1998; Callen et al., 2003; O’Regan & Oster, 2005;
Gu et al., 2010; Agoraki et al., 2010).
Yermack (1996), in a review of the earlier work of Monks & Minow (1995), argues that large boardrooms tend to
be slow in making decisions, and hence can be an obstacle to change. A second reason for the support for small board size
is that directors rarely criticize the policies of top managers and that this problem tends to increase with the number of
directors.
Examining the relationship between board size and firm performance, they conclude that the smaller the board
12 Chris Ogbechie & Bongo Adi

size the better the performance, and thus proposed an optimal board size of ten or fewer. John & Senbet (1998) maintain
that the findings of Yermack have important implications, because they may call for the need to depend on forces outside
the market system in order to determine the size of the board.
Eisenberg et al. (1998) also find a negative correlation between board size and return on assets and operating
margin for a sample of 900 small and mid-sized finish firms. And Brown & Maloney (1998) find that larger board size
predicts lower stock price returns to acquiring firms intake overs. However, Bhagat & Black (1998) find that the inverse
correlation between board size and performance is not robust to the choice of performance measure.
Guest (2009) examine the impact of board size on firm performance for a large sample of 2,746UK listed firms
over 1981-2002. They find that board size has a strong negative impact on profitability, Tobin’s Q and share returns, which
is robust across econometric models that control for different types of endogeneity. He finds no evidence that firm
characteristics that determine board size in the UK lead to a more positive board size–firm performance relation.
In contrast, he finds that the negative relation is strongest for large firms, which tend to have larger boards. His evidence
supports the argument that problems of poor communication and decision-making undermine the effectiveness of large
boards.
Similarly, Eyenubo (2013) examined the relationship between bigger board and financial performance, by looking
at sample of 50 firms listed in the Nigerian Stock Exchange during the period 2001–2010. Using regression analysis, his
finds that bigger board Size affects the financial performance of a firm in negatively.
In all, the theoretical and the empirical supports for agency theory of board size have become popular.
The argument points towards a smaller board size for optimal performance. But what does “small” imply in the context of
Nigeria? To what extent must board size grow before it becomes large, thereby impeding firm performance? Of course the
agency theory does not answer the question, which warrants a further exploration of the board size discourse.
Resource Dependency Theory
In direct contrast to the agency theory, the resource dependency theory suggests that companies with larger boards
do better. Large boards are more likely to contain directors with greater diversity in education and industry experience.
This diversity allows the board members to provide management with high quality advice (Zahra & Pearce, 1989).
The argument is that each new board member brings both expertise and access to resources. Thus, having more board
members would provide the firm with greater expertise and access to resources.
These resources could include access to markets, access to new and better technologies, and access to raw
materials among other things (Ning et al. 2010). According to Dalton & Dalton (2005), the advantage of larger board size
is the greater collective information that the board subsequently possesses and hence larger boards will lead to higher
performance.
Proponents of resource dependence theory argue that organizational survival is dependent on the ability to access
critical resources from the environment (Casciaro and Piskorski, 2005; Pfeffer and Salancik, 1978). Firms actively manage
their resource environments by maintaining external linkages to organizations on which they depend for critical resources
(Hillman, 2005; Westphal et al., 2006). Boards also react in the same way, for example by adding a representative of a
critical resource to the board constitutes a way of managing this dependence and benefiting the firm (Ogbechie, 2012).
The argument is that directors with high social capital can bring information about the external environment, other firms’
strategies, and prospective managerial talent to the firm (Certo, 2003; Davis, 1999; Haunschild, 1993). Ultimately, these
Exploring Board Size and Firm Performance: Is there a Kuznets’ Curve Relationship? Evidence from Nigeria 13
ties can impact on the performance of the board and hence of the firm (Nahapiet & Ghoshal, 1998). Modern boards are
therefore composed on the basis of the resources the directors will bring to the board (Ogbechie, 2012)
There is also empirical support for the resource dependency theory. Pfeffer (1973) finds that board size is
positively correlated with the sources of funding in health care firms. Provan (1980) finds a positive and significant
correlation between board size and nonprofit firms access to sources of funding. And Chaganti et al. (1985) find that firms
with smaller boards are associated with a higher rate of bankruptcy.
More so, Booth & Deli (1996) contend that environmental uncertainty generally leads to large board size, arguing
that larger board allow firms access to the expertise necessary to overcome this uncertainty. Accordingly, in a
meta-analysis of 131 studies drawn from an aggregate of 20,620 companies, Dalton et al. (1999) find a non-zero and
positive board size and performance relationship. And Kiel & Nicholson (2003) find positive board size effects in
Australia.
Kyereboah-Coleman & Biekpe (2005) find a positive relationship between the board size and the value of firm in
developing markets. And Eyenubo (2013) maintain that a bigger board brings higher management skills and makes it
easier for the board to make strategic decisions that result in improving the value of a firm. Similarly, a CEO can easily
manipulate a smaller board and can compromise the efficiency and independence of a board. But larger boards are more
independent and efficient, as the CEO cannot manipulate it.
Notwithstanding these findings and the alluring theoretical position of the resource dependency proponents,
common sense suggests that much as larger board size may be beneficial to firm performance, boards cannot get infinitely
large, for then it creates more costs than benefits to the firm. So, there is the need for knowledge on how large board size
needs to be before further increase becomes suboptimal. And this cannot be addressed in the context of just the resource
dependency theory.
Agency and Resource Dependency Theories: Is there a Kuznets’ Curve?
In view of optimality and the recognition that there could be a mid-way between having too large or too small
board sizes, Hillman & Dalziel (2003) integrates agency theory and resource dependency theory. Based on the idea that
directors serve as monitors and provide access to resources, they argue that viewing governance through only one theory
provides an incomplete understanding of board functions and small boards may not be optimal for all companies.
Other studies also find that small boards may not always be optimal. Several prior studies have provided support for
optimal board size.
Hiner (1967) argues that there is a threshold where board size may have an inverse impact on firm performance.
Lipton & Lorsch (1992) argue that the optimal board size appears when additional benefits, such as additional information,
expertise, and monitoring of more board members is equal to additional costs such as free-riding problems, operational
complexity and member conflicts within the board. Yermack (1996) finds that the correlation between firm value and
board size has a convex shape, suggesting that the largest proportion of firm value lost, due to the increase in board size,
occurs when a board grows from small to medium size. And Sanda et al. (2010) find that the relationship between firm
performance and board size is positive up to a point. Beyond this point a negative relationship is predicted to set in.
In other words, there is the possibility of an optimal board size in Nigeria, for size matters. Following these empirical
positions, it can be deduced that the debate is no more on small or large boards. But on how small or large board size need
to be for optimality to be attained. The argument goes that as board size rise it adds more value to the firm, but then
becomes detrimental to the firm beyond an optimal point.
14 Chris Ogbechie & Bongo Adi

This relationship between board size and firm performance, which allows for optimality, typically replicates what
is popularly termed the Kuznets’ curve – inverted u-shaped curve of inequality – in economic development theory.
Kuznets (1995) used this hypothesis to show that as economies develop, they first witness rising levels of inequality, then
attain a peak where inequality is at its maximum, and then begin to decline as development rise.
Going forward, it follows that such relationship established by Kuznets, which we experiment in board size in
Nigeria (see figure 1 below), has been empirically argued indirectly in various climes, as recorded in the preceding
paragraphs.

Figure 1: Kuznets’ Curve of Board Size
The curve above represents what Hiner (1967), Lipton &Lorsch (1992), Yermack (1996), Hillman &
Dalziel (2003) and Sandaet al. (2010) argued in the preceding paragraphs. The upward sloping portion of the curve
represents the argument of the resource dependency theorists, while the downward sloping portion represents the argument
of the Agency theorists. We proceed to methodologically experiment the existence of this Kuznets’ curve – a mid-way
between the Agency theory and Resource dependency theory – in board size in the Nigerian context.
3. THEORETICAL FRAMEWORK
In view of the shortcomings of considering board size from the perspective of a single theory and the possibility
of the existence of a Kuznets’ curve of board size, this study is based on the combination of both the agency and resource
dependency theories. Thus we adopt the approach of Hillman & Dalziel (2003), Ning et al. (2010) and integrate agency
theory and resource dependency theory to explore what the optimal board size of corporations is in Nigeria. Agency theory
outlines the costs of large boards, while the resource dependency theory offers that there are benefits of having a large
board. Therefore, if both theories are reliable, an optimal board will be one that strikes a balance – minimizing the costs of
too small or too large boards, and maximizing benefits. This balance is what we explore in Nigeria in the context of the
Kuznets’ inverted u-shaped curve shown in figure 1 above.
Data and Methodology
We use a sample of 100 firms quoted in the Nigerian Stock Exchange during the period 1999 – 2008. In achieving
Exploring Board Size and Firm Performance: Is there a Kuznets’ Curve Relationship? Evidence from Nigeria 15
the first objective of this study, we explore the possibility of a mean reversion trend3 in board size in Nigeria, given the
sample. Here we categorize the board sizes into the following groups: 3-5, 6-8, 9-11, 12 and above, and compare the mean
of these board size groups in 1999 and in 2008, subjecting the differences to tests of significance to see if there is
movement towards a given board size range, which could then be the optimal size. The sample is further splitin to
two – financial and non-financial firms – because of different regulatory codes4 that apply to these sectors, which affect
board size. This is necessary to see if the earlier finding will be sustained. A random effect model (subject to Hausman test
for model selection) including firm performance, board size, squared board size and relevant controls – Debt, firm size
(measured by total asset), age and number of outside directors – is further employed to calculate optimal board size value
and for comparison with findings from the previous method. These methodological approaches will reveal what board size
is in Nigeria at the optimal point shown in the Kuznets’ curve in figure 1 above.
For the second objective, dynamic linear panel-data model is adopted. A major econometric concern highlighted
in recent studies is that past and current firm performance determines board size (Lehn et al., 2004; Boone et al., 2007;
Coles et al., 2008; Guest, 2008; Linck et al., 2008; Guest, 2009; Sanda et al., 2010). Ordinary least squares and fixed
effects analysis will be biased in the presence of such dynamic and simultaneous endogeneity (Guest, 2009). Therefore, in
order to deal with the problems of unobserved heterogeneity, simultaneous endogeneity and dynamic endogeneity, we
employ the optimal Generalized Method of Moment (GMM) dynamic panel data estimation (Arrelano-Bond dynamic
panel-data estimation), as found robust by various studies (Arrellano & Bond, 1991; Wintoki, 2007; Guest, 2009).
This approach allows board size to be determined by past and present performance, and therefore provides an appropriate
econometric specification fordealing with the endogeneity problems. The approach includes lagged performance as an
explanatory variable, and takes first differences, which eliminates thecompany specific fixed effects. GMM estimation then
uses lagged levels (by two periods and earlier) of board size and performance as instruments, which controls for both
dynamic and simultaneous endogeneity. Debt, firm size, age and number of outside directors are used as controls and are
also treated as endogenous. And two measures of firm performance –Return on Asset and Return on Equity – are used as
dependent variables.
4. RESULTS AND DISCUSSIONS
Does Kuznets’ Curve of Board Size Exist in Nigeria?
In answering the above question, which is the major thrust of this study, we examined the possibility of mean
reversion trend in board size in Nigeria over time. We divided the board sizes of all the 100 firms sampled in this study
into four groups – 3-5, 6-8, 9-11, 12 and above – and computed the mean of these groups in 1999 and in 2008, subjecting
their differences to tests of significance. We first employed the Wilcoxon-Mann-Whitney test, relaxing the assumption of
normal distribution in the board size sample being examined. The result of the various tests for each of the board size
groups (see appendix 1 table 1a through 4a) showed no significant difference between the means of board size groups
in 1999 and in 2008. For the groups 3-5, 6-8, 9-11 and 12 and above directors, the probability of the z-score is 0.3782,
0.5536, 0.3494 and 0.5905 respectively, which are all nonsignificant at the 0.05 level of significance.

3Ning et al (2010) used the term mean reversion trend to explain the finding amongst a sample of US firms between 1988
and 1999 that firms with fewer than 8 – 11baord size tend to significantly increase their board size towards this
range,whereas firms with board size larger than this range tend to reduce it.
4 The code of Best Practices for Public Companies, developed by the Securities and Exchange Commission (SEC),
recommends that the Board of Directors shall be composed of both executive and non-executive directors and should not
exceed 15 persons or be less than 5 persons in total. Whereasthe Code of Corporate Governance for Banks in Nigeria,
developed by the Central Bank of Nigeria (CBN) recommends a board size of between 5 and 20 directors.
16 Chris Ogbechie & Bongo Adi

We re-examine this finding by upholding the assumption of normal distribution in the board size sample and thus
apply t test on the board size group mean differences. The results captured in appendix 1 table 1b through 4b show that the
mean difference of all the board size groups 3-5, 6-8, 9-11 and 12 and above directors, with probability values 0.8169,
0.5722, 0.5205 and 0.7527 respectively, are again statistically nonsignificant at the 0.05 significance level. This indicates
that over time, board of directors in Nigerian firms does not move towards a given value which could be judged the
optimal size. Although Ninget al (2010) used this same approach and found optimal board size of 8 – 11 directors for US
publicly quoted firms, we do not find such evidence for Nigerian firms in this study.
In a bid to verify the robustness of this finding, we further employ a random effect model, with board size,
squared board size, number of outside directors, debt, firm size and firm age as the explanatory variables for firm
performance (return on asset, ROA). The motivation for this is to use partial differential calculus to compute optimal
values for board size. We first subject this model to a Hausman test for fixed effects. The result of this test
(see appendix 4 table 13), with a probability value of 0.5091, which is nonsignificant at 5% level of significance, leads to
the acceptance of the null hypothesis that a random effect model is appropriate and will yield consistent estimates.
We therefore estimate the random effect model, with results (see appendix 4 table 12) showing both board size
and squared board size as nonsignificant explanatory variables for firm performance (ROA) in Nigeria. Moreover, Board
size is found to have positive impact on firm performance. This impact turns negative when board size is squared, implying
that there exists Kuznets’ curve of board size in Nigeria. Specifically, the implication of this is that at some point, further
increase in board size no longer accounts for improved firm performance as the resource dependency theorists argue.
It rather begins to constitute a bane to organizational performance. In other, when boards are unnecessarily large, they
become unwieldy. Decision making becomes slow. Issues of moonlighting erupt. And the board may suffer executive
capture. All these then lead to poorer organizational performance as posited by the agency theorists.
Notwithstanding the established evidence of Kuznets’ curve in board size in Nigeria above, we, however, do not
proceed to compute the optimal value of board size using partial differentiation. This is because the impact of both board
size and squared board size variables on firm performance are nonsignificant. This, in essence, validates the earlier
evidence that there is no specific board size that could be considered as optimal for all the publicly quoted companies in
Nigeria. However, the evidence of Kuznets’ curve of board size in Nigeria implies that firms operating within different
institutional and geographical settings in the country need to work out what there optimal board size is. Not doing this will
warrant firms reaping a sub-optimal value from their board of directors, or a beyond optimal board value which will still be
lower and inefficient.
In view of the different regulatory codes that govern financial and non-financial publicly quoted firms in Nigeria,
we consider further split the sampled 100 firms into financial and non-financial firms. We reapply the above tests of mean
reversion trend to these two categories. In the nonfinancial firms, the Wilcox on-Mann-Whitney test
(see appendix 2 tables 5a through 8a) for board size groups 3-5, 6-8, 9-11 and 12 and above directors has probability values
0.3782, 0.6153, 0.0958 and 0.6872, which shows evidence of no significant difference in the means of these groups in
1999 and 2008 at the 0.05 significance level. The t tests for these same groups (see appendix 2 tables 5b through 8b)
uphold this result, thereby corroborating the earlier evidence of non-movement towards a perceived optimal board size.
Interestingly, the case of financial firms in Nigeria differs from the earlier findings. First, none of the 31 financial
firms sampled in the study has board size falling in the group 3–5 directors. Subjecting the mean difference of remaining 3
groups to tests of significance, both the Wilcoxon-Mann-Whitney test and t test, with probability values of 0.0423 and
Exploring Board Size and Firm Performance: Is there a Kuznets’ Curve Relationship? Evidence from Nigeria 17
0.0325 respectively, show statistically significant difference in the mean of board size group 6 – 8 directors in 1999 and
2008. For the other two groups, 9 – 11 directors and 12 and above directors, there is no significant difference
(see appendix 3 tables 9a to 11b).
The board size group, 6-8 directors, with significant mean difference has its mean increase from 6.43 in
1999 to 7.38 in 2008. The board size group 12 and above directors is the most stable group, as its mean fell from 14.75 in
1999 to 14.53 (a difference of 0.22) in 2008, whereas that of the group 9-11 directors fell from 10.08 in 1999 to 9.8
(a difference of 0.28) in 2008.
The implication here is that the most stable and hence optimal board size for financial firms in Nigeria is at least
12 directors. Movement of board sizes over time, however, goes to emphasize the relevance of both the agency theory and
the resource dependency theory in board size choice among financial firms in Nigeria. We observe that financial firms in
the smaller board size group (6-8 directors) raised their board sizes in line with the resource dependency theory argument
of benefits of large boards, which comes from diversity, experience, social capital and access to resources
(see Pfeffer & Salancik, 1978; Haunschild, 1993; Nahapiet & Ghoshal, 1998; Davis, 1999; Certo, 2003; Casciaro and
Piskorski, 2005; Dalton & Dalton, 2005; Ogbechie, 2012). On the other hand, in the larger board size categories, financial
firms’ board size decisions rely more on the agency theory. They follow the agency theory argument that substantial
increase of the board size results in a slowdown in decision making and an increase in costs, leading to lower
firm/organizational performance (Alexander et al., 1993; Eisenberg et al., 1998; Callen et al., 2003; O’Regan &
Oster, 2005; Gu et al., 2010) and reduce board size over time as observed.
Capping up, we infer that neither the agency theory nor the resource dependency theory on its own explains firms’
board size decision in Nigeria. Firms act in line with both theories at different times in the organization’s life.
This submission supports the conclusion of Hillman & Dalziel (2003); Ning et al (2010) and Sanda et al. (2010) who assert
that organizations consider agency costs and value maximization when making board size decisions. Boardsoften weigh
the costs and benefits of a large or small size and choose the size that they deem optimal.
Does Size Matter?
Various attempts have been made in the past to assess the role of board size in organizational performance.
These efforts have often been faulted due to the negligence of endogeneity problems in the methodological approach of
these studies. Simultaneous endogeneity and dynamic endogeneity are often highlighted to render the results of various
studies on the impact of board size on firm performance suspect. In an attempt to circumvent these endogeneity concerns
and produce consistent and reliable results, we resorted to dynamic panel data instrumental variable approach. We thus
employed the Arrelano-Bond dynamic panel data estimation technique. In using this method, we modeled board size
against two measures of firm performance – Return on Assets (ROA) and Return on Equity (ROE) – and used debt,
number of outside directors, firm age and firm size (total assets) as control variables. Appendix 5 table 14a shows the result
of this estimation for ROA. In this result we controlled for heteroskedasticity and found the coefficient of board size to be
positive but no nsignificant at the 0.05 significance level. We confirm the efficiency of this result by testing for the
presence of autocorrelation in the residuals. Table 14b in appendix 5 shows the result of autocorrelation tests to the 4th
order. It reveals the absence of autocorrelation in the result, implying that our finding with respect to the impact of board
size on ROA in Nigeria is efficient.
We repeat the above estimation procedure using ROE as the measure of firm performance. Again, after ensuring
that the estimation result is heteroskedasticity consistent, we find the coefficient of board size to be positive and no
18 Chris Ogbechie & Bongo Adi

nsignificant (see appendix 5 table 15a). We further test for autocorrelation to the 4th order, and the result
(see table 15b appendix 5) reveals the absence of autocorrelation in the residuals. This resultsupports the work of
Pfeffer (1973), Kiel & Nicholson (2003), Kyereboah-Coleman & Biekpe (2005), Sanda et al. (2010) who found board size
to have positive impact on firm performance. It however contradicts the work of Gertner & Kaplan (1996), Denis &
Sarin (1999), Ning et al (2010) and Eyenubo (2013) whose studies support the agency theory that rising board size
significantly impedes firm performance.
It therefore follows that board size does not significantly impact organizational performance in Nigeria.
Connecting this with the inferences in the previous section, we surmise that before board size reaches the optimal point,
increasing the size of an organization’s board improves firm performance in Nigeria, but only marginally. This negligible
contribution to firm performance however disappears and even turns negative beyond the optimal board size of the firm in
question. Additionally, Nigerian firms seeking for ways to improve their financial performance need not place much
emphasis on increasing board size, as this has been found to play only a marginal and negligible role. Other organizational
and institutional factors are more crucial for firm performance in the country. These factors, which are subject to
verification and opens the door for further research, could be organizations’ management style, organizational culture,
employee remuneration and motivation, economies/diseconomies of agglomeration and scale and economic geography.
5. CONCLUSIONS
The numerous corporate governance failures in different parts of the world in recent years have renewed the
interest of both researchers and regulators on corporate governance efficiency. Board of directors is understood to be a very
essential component of the corporate governance structure. Boards are the managers’ manager.
The extent to which boards of directors are able to manage the managers is defined, in the language of the agency
and resource dependency theories, by size. This study therefore sought to underline the optimal board size for publicly
quoted companies in Nigeria. This optimal board size if it exists ought to be the most efficient for companies, in terms of
contribution to corporate governance an organizational performance.
We assessed the possibility of firms moving towards a particular board size over the period of the study. Evidence
for all the 100 sampled firms show that there is no significant movement towards a given board size. However, this finding
does not hold when the sampled firms are divided into financial and nonfinancial firms. For the financial firms, we find
evidence that the target board size group is at least 12 directors. Furthermore, a random effect model showed evidence of
an inverted u-curve of board size for all the sampled publicly quoted companies in Nigeria.
Thus although we are unable to establish an optimal board size for the public quoted companies as a whole, we
are able to establish that such an optimal size exists for each firm, given the evidence of Kuznets’ curve of board size. It is
not surprising that a unique optimal board size could not be established for all the publicly quoted companies, because
these firms are actually of different sizes and operate in often divergent institutional, regulatory and organizational
contexts. Given this,firms therefore need to consider there geographical, organizational and institutional specificities, then
establish what an optimal board size is for themselves. This size does not need to be static; it may have to change as the
institutional, organizational and operational realities of the firm changes.
The existence of Kuznets’ curve of board size in Nigeria further clarifies that the impact of board size on firm
performance in Nigeria neither supports only the agency nor resource dependency theory. Both theories play a role.
The finding from the optimal GMM estimation and the random effect model both affirms this position. Notwithstanding
this, how large or small board size is does not significantly influence organizational performance.